learning-dynamics

Tag

Cards List
#learning-dynamics

Pseudospectral Bounds for Transient Amplification in Coupled Gradient Descent

arXiv cs.LG · 6d ago Cached

This paper develops a sharp pseudospectral theory for block-triangular Jacobians in coupled gradient descent, proving Kreiss-constant bounds and establishing iteration complexity results. The work exposes non-asymptotic, instance-dependent transient amplification phenomena relevant to bilevel optimization, two-time-scale stochastic approximation, and GAN training.

0 favorites 0 likes
#learning-dynamics

Support Before Frequency in Discrete Diffusion

arXiv cs.LG · 2026-05-15 Cached

This paper proposes the 'support-before-frequency' hypothesis for discrete diffusion models, suggesting that models first learn the support (admissible sequences) before refining frequencies within the support. Theoretical analysis of small-noise reverse kernels and experiments on masked language diffusion models support this claim.

0 favorites 0 likes
#learning-dynamics

State-Space NTK Collapse Near Bifurcations

arXiv cs.LG · 2026-05-14 Cached

This paper develops a local theory of gradient descent near bifurcations in dynamical models, showing that the state-space neural tangent kernel collapses to a rank-one operator that dominates learning dynamics, making optimization effectively low-dimensional and predictable from normal forms.

0 favorites 0 likes
← Back to home

Submit Feedback